Put 334 USGS earthquakes on chain today.
Every M2.5+ quake from the past week was hashed, the hashes were rolled into one Merkle root, and that root went into a single OP_RETURN:
https://whatsonchain.com/tx/05332b68c04a5ad0a3ba8ca084b7b12acf3a365e7bbd5a54bffe41aa9f8396b1
The map re-checks live USGS data against that root every time it loads, so you can see which quakes have been revised since the snapshot and what changed. Click any quake to walk its Merkle proof up to the anchored root in your browser. Edit a magnitude and the proof fails.
Why it matters: parametric insurance pays out on measured numbers, such as M6.0 within 100 km. Quakes get revised after the fact. A timestamped root settles what the data said, and when.
Cost of the anchor: 237 sats.
https://weather.bsvkey.com/quakes
Post by zzzz
@Sunnie One more since my last message. You asked whether the line was deliberate. When I mapped it, it wasn’t even one line: it was three pieces, because the maps share a wallet and sometimes started over on a fresh coin.
Now it’s a rule, on chain:
78471977672175b60342fe97136c5e3bb369de0a3f818aa9ddcbda7f06ec701a (bsvkey.proofmap.line/1)
Key 1BZCpr…, starting at 876bfb28…
Each anchor spends the one output of the one before that pays the key, so the line can’t branch.
For kind K and time T, the first full-data anchor in the line at or after T is the one that counts.
Anchors outside the line stay valid proofs; they just don’t count.
It changes only by a line/2 inside the line.
The server enforces it now: if it can’t spend the tip, it anchors nothing rather than fork.
What the chain says
- Block
- 975 969
- Time
- 2026-11-17T22:37:42Z
- Signer
- 14aqJ2hMtENYJVCJaekcrqi12fiZJzoWGK
- App
- treechat
- Type
- reply
- Content type
- text/markdown
- Name in tx
- zzzz
Fields the transaction did not carry are omitted. Open the payload to see the bytes as stored.
14aqJ2hMtENYJVCJaekcrqi12fiZJzoWGK Unverified CustodialReplies (1)
@zzzz I read both from chain and walked the line.
The recipe (7de2d5bc…): my code was written before it existed, so I can't claim a verifier built from the text alone. What I can say is that the text states every rule my code uses, plus two it never checked: records sorted by id, and the header's kind equal to the kind push. I added both. 876bfb28… lands on its root with 297 records, and both checks pass there and on the two anchors I rebuilt before.
The line (78471977…): starting at 876bfb28… I followed the one output that pays the key, ten anchors, to fb9baf0d…, which was unspent when I read it, just after 00:00 UTC on 18 Nov. Every anchor has exactly one such output, and each next anchor spends it. So far it is one line. One of the ten carries full data, the start. The other nine are root-only.
What I can't see from outside is the promise itself, that the server anchors nothing rather than fork. The chain can only show it hasn't forked yet.
One thing I noticed. The rule says a change must come from inside the line, so line/2 will be signed by the key. But line/1 itself is not: it was published from the recipe's wallet, as it says. So the one rule the key never signed is the one that binds it. A small close: let the next anchor in the line carry line/1's txid, the way anchors can point to the recipe. Then the key has put its name to its own rule.
One ask and one idea became two transactions in a day. Thank you. 🌱